untitled Abstract Background. HIV-associated focal brain lesions (FBLs) are caused by opportunistic infections, neoplasms, or cerebrovascular diseases. In devel- oped countries toxoplasma encephali- tis (TE) is the most frequent cause fol- lowed by primary CNS lymphoma (PCNSL). Guidelines based on these causes have been proposed and suc- cessfully implemented. These guide- lines do not consider the causes of HIV-associated FBL in developing countries where treatable infections predominate. The guidelines as pro- posed would adversely influence out- come if applied to patients in develop- ing countries. Objective. To determine a practical approach to the management of HIV- associated FBLs in developing coun- tries. Design. Prospective case series. Methods. Management based on pre- sumed aetiologies of the FBLs deter- mined by collating information obtained from computed tomogra- phy (CT) scans, cerebrospinal fluid and blood studies, concurrent non- neurological illness and response to treatment. Results. The principal presumed cause of FBL was tuberculosis (TB) (69%), and the therapeutic response was good in 69% of patients. Conclusions. In developing countries infections are the predominant cause of HIV-associated FBL. The principal cause is determined by the infection that is endemic to the population being studied. Empiric treatment based on limited investigations should be directed according to the nature of this infection. A modified algorithm is proposed. Introduction Focal brain lesions (FBLs) caused by opportunistic infections, neo- plasms, or cerebrovascular diseases are common neurological consequences of HIV infection.1 In developed countries toxoplas- ma encephalitis (TE) is the most fre- quently identified cause of HIV-asso- ciated FBL, followed by primary CNS lymphoma (PCNSL).1,2 On this basis the Quality Standards Subcommittee of the American Academy of Neurology (AAN) proposed that HIV-positive patients presenting with FBL should be empirically treated for toxoplasmosis initially.1 Failure to improve clinically or radiologically over the succeeding 10- to 14-day period warrants stereotactic brain biopsy in order to institute specific and appropriate therapy.1 An alterna- tive non-invasive approach that is oth- erwise similar is based on the fact that brain biopsies do not influence sur- vival in these patients.2 With this approach diagnosis is determined by the response to anti-toxoplasma treat- ment.2 Failed responders are reviewed and presumed to have PCNSL or tox- oplasmosis resistant to standard ther- apy and treated appropriately.2 These guidelines and approaches to the management of HIV-associated FBL have been effective in developed countries. In developing countries infections are the main cause of FBL associated ORIGINAL ARTICLE 5 SA JOURNAL OF RADIOLOGY • July 2005 Management of HIV-associated focal brain lesions in developing coun- tries – experience at Chris Hani Baragwanath Hospital M Modi FCRad (SA), MMed Department of Radiation Sciences University of the Witwatersrand Johannesburg A Mochan FCP (SA) G Modi PhD (Lond), FCP (SA), FRCP (Lond) Department of Neurosciences University of the Witwatersrand Johannesburg 5-13pgs 9/28/05 9:20 AM Page 5 with HIV.3-10 Studies from African and Central American countries describe either tuberculosis (TB) or toxoplas- mosis as the two most frequently occurring causes.3-8 These observa- tions have been corroborated in simi- lar studies from India.9,10 The relative frequencies with which these infec- tions cause FBL are not well docu- mented. The nature of the HIV-asso- ciated FBL is probably determined by the infection/s that are endemic to the population being studied.7 TB is endemic in developing countries.11 This therefore implies that if the AAN or British guidelines are strictly applied to developing regions where TB is endemic large numbers of patients will be inappropriately man- aged. We therefore prospectively studied and managed 32 patients with HIV- associated FBL. Our approach was to presume the most likely cause in our setting determined by collating infor- mation obtained from computed tomography (CT) scans, cere- brospinal fluid (CSF) and blood stud- ies, concurrent non-neurological ill- ness and response to treatment. We did not emphasise the need for accu- rate histological diagnosis of the FBL. Study design This was a prospective case series. Patients and methods We studied 32 HIV-infected patients (older than 18 years) with FBL as diagnosed on CT scans. The study was carried out at the Chris Hani Baragwanath Hospital (CHBH) in Soweto, South Africa. Patients were recruited over a 12-month period. CHBH is a 3 300-bed public universi- ty hospital that serves a predominant- ly black urban population of approxi- mately 3 million people. The patients were all inpatients, admitted to the medical wards of the CHBH. All patients in the study were black, heterosexual, non-intravenous drug abusers and were anti-retroviral therapy-naive. None of the patients was on prophylactic treatment for Pneumocystis carinii or toxoplasma. The following data were recorded. Demographics: Age, sex, ethnic group, medication/drug use, and associated non-neurological illnesses. Blood: Full blood count, erythro- cyte sedimentation rate (ESR), glu- cose, urea and electrolytes, serum cal- cium, phosphate and magnesium, liver function tests, T-cell subsets, HIV viral loads (HIV-1 RNA), serological tests for syphilis (WR, TPHA, FTA- Abs), cysticercosis (ELISA), Toxoplas- ma gondii (CFT), typhoid (TMX), viral studies for hepatitis A, B and C, cytomegalovirus (CMV), HTLV-I and blood cultures. Cerebrospinal fluid (CSF): Chemistry, cell counts, cytology, and adenosine deaminase (ADA) level. TB ELISA and polymerase chain reaction (PCR), syphilis, cysticercosis and tox- oplasma serology, HIV viral load (HIV-1 RNA), PCR for herpes sim- plex, varicella zoster, cytomegalovirus (CMV), Epstein-Barr virus (EBV) and herpes simplex type 6, India ink staining, cryptococcal antigen, bacter- ial (including Mycobacterium tubercu- losis) and fungal cultures. Sputum: Cytology, microscopy, culture and sensitivity, and acid-fast bacilli (AFB). Radiology: Chest X-rays (CXRs). CT scans without and with contrast enhancement were done prior to treatment and repeated days or weeks later depending on the patients’ clini- cal condition. Presumed diagnosis of FBL: The accurate and definitive diagnosis of FBL is dependent on histopathology. In this study as we were unable to obtain histopathological diagnosis we presumed the diagnosis of the FBL by collating information obtained from the CT scans (at the time of presenta- tion and on empiric treatment), CSF and blood studies, CXR findings, associated non-neurological illness/es and response to treatment. Presumed diagnosis of TB was based on some or all of the following: (i) CT brain scan appearance, i.e. hypodense or isodense rounded lesions with irregular walls of varying thickness, oedema and mass effect, cortical location, ring or nodular enhancement, and increased basal meningeal enhancement; (ii) positive TB culture in the CSF; (iii) pul- monary TB (PTB) on CXR and/or AFB on sputum microscopy; (iv) increased protein, decreased glucose, pleocytosis (lymphocytes and/or polymorphonuclear cells), positive TB ELISA, PCR in the CSF; (v) nega- tive toxoplasma CFT; and (vi) response to TB treatment, clinically and radiologically. Neurosyphilis was diagnosed on the basis of CSF serology (WR, TPHA, FTA). Presumed diagnosis of toxoplas- mosis was based on some or all of the following: (i) positive toxoplasma CFT; (ii) CD4+ count less than 100 cells/mm3; (iii) hypodense, multiple (> 5) rim-enhancing lesions with oedema, with basal ganglia and ORIGINAL ARTICLE 6 SA JOURNAL OF RADIOLOGY • July 2005 5-13pgs 9/28/05 9:20 AM Page 6 grey/white matter zones as common sites; and (iv) response to toxoplasma treatment. Diagnosis of neurocysticercosis was based on some or all of the fol- lowing: (i) regular thin-walled, cystic, ring-enhancing lesions (often with a scolex); (ii) intracranial calcifications; and (iii) positive blood and/or CSF serology (ELISA). Cryptococcosis was diagnosed on CSF studies (India ink, latex aggluti- nation, fungal cultures). Presumed diagnosis of PCNSL was based on some or all of the fol- lowing: (i) hyperdense lesion/s, often confluent, with no oedema, homoge- neous contrast enhancement, peri- ventricular/subependymal in loca- tion, corpus callosum involvement; (ii) CD4+ count less than 100 cells/mm3; (iii) atypical lymphocytes in the CSF; (iv) EBV PCR-positive in CSF; and (v) no response to anti-TB or anti-toxoplasma treatment. Presumed diagnosis of progres- sive multifocal leucoencephalopathy (PML) was based on some or all of the following: (i) lesions confined to white matter, hypodense, non- enhancing appearance on CT scan, T2 hyperintense and T1 hypointense sig- nal on MRI scans if performed, no mass effect; and (ii) JC (John Cunningham) virus PCR-positive in CSF. Treatment: Patients were treated with specific medication based on the presumed diagnosis: TB (isoniazid, rifampicin, pyrazinamide, ethambu- tol), neuroyphilis (IVI penicillin G), toxoplasmosis (sulfadiazine, pyrime- thamine), and cysticercosis (albenda- zole). If response was good, the initial treatment regimen was continued. If response was poor, a new or second presumed diagnosis was entertained, and additional treatment for this diag- nosis was instituted. The additional or new treatment was as described above. A good response was one in which the patient showed ongoing improve- ment measured radiologically by a reduction in size and/or number of lesions on scan. A poor response was one in which the patient showed no improvement with no change or worsening of the FBL on scan. Results (Tables I - III) Age and sex The age range was 18 - 49 years (mean 33.5 years), with 19 male patients and 13 female patients (male/female 1.5:1). Clinical presentation On CNS examination there were focal signs in 21 patients (66%), nor- mal in 9 patients (28%), dementia in 6 patients (19%), and encephalopathy in 2 patients (6%). Patients were examined clinically on admission by the authors GM or AM who were blinded to the CT scan result. Once patients were recruited into the study the follow-up examina- tions done by either GM or AM were not blinded. Non-neurological illnesses Non-neurological illnessess were present in 27 patients (84%): active PTB (19 patients (60%)), active PTB and tuberculous breast abscess (1 patient), tuberculous lymphadeni- tis (1 patient (diagnosed on fine nee- dle aspirate)), Pneumocystis carinii pneumonia (PCP) (1 patient (diag- nosed on sputum cytology)), and pos- itive syphilis blood serology (WR, TPHA) (11 patients). CD4 counts and staging Staging was determined using the Centers for Disease Control (CDC) 1993 revised classification system for HIV infection and AIDS-defining ill- nesses.12 The patients had CD4+ T-lym- phocyte counts ranging from 1 to 768 cells/mm3. Four patients (12%) had CD4+ counts of greater than 500 cells/mm3. Four patients (12%) had CD4+ counts between 200 and 500 cells/mm3. Twenty-four patients (76%) had CD4+ counts of less than 200 cells/mm3. HIV viral loads HIV blood viral loads were obtained in 29 of the 32 patients. These ranged from 10 750 copies/ml to > 750 000 copies/ml. CSF results There was elevated CSF protein, with or without pleocytosis in 23 patients. Cysticercosis ELISA was pos- itive in 8 patients, TB ELISA was pos- itive in 1 patient, TB PCR was positive in 1 patient, TB culture was positive in 1 patient, cryptococcal antigen was positive in 1 patient, syphilis serology was positive in 1 patient, atypical B- cells were present in 1 patient, and toxoplasma immunoglobulin G (IgG) was positive in 1 patient, normal in 7 patients (22%), and not performed in 2 patients for practical reasons. Radiology Chest X-rays Features of PTB were present in 20 patients (63%), there were bilateral peri-hilar reticulo-nodular infiltrates in 1 patient, and CXR was normal in 11 patients (34%). ORIGINAL ARTICLE 7 SA JOURNAL OF RADIOLOGY • July 2005 5-13pgs 9/28/05 9:20 AM Page 7 CT scans These were performed on admis- sion and repeated days or weeks later depending on the patient’s clinical condition. The following observations were made: 12 patients had solitary lesions (38%) and 20 patients had multiple lesions (62%). In the 12 patients with solitary lesions the CT scans were suggestive of TB in 8 patients, neurocysticercosis in 3 patients and PCNSL in 1 patient. In the 20 patients with multiple lesions the CT scans were suggestive of TB/toxoplasmosis (12 patients), neu- rocysticercosis (3 patients), PML (1 patient), and TB/toxoplasmosis and neurocysticercosis (2 patients). Two patients had cerebral infarcts. Meningeal enhancement (Fig. 1) was present in 5 of the 12 patients with solitary lesions and in 12 of the 20 patients with multiple lesions. None of the patients diagnosed radio- logically as neurocysticercosis, PCNSL or PML had meningeal enhancement. The radiological diagnoses were therefore TB (17 patients), neurocysticercosis (6 patients), PML (1 patient), PCNSL (1 patient), TB/toxoplasmosis and neurocysticer- cosis (2 patients), TB/toxoplasmosis (3 patients), and in 2 patients cerebral infarcts (both had basal meningeal enhancement). Presumed causes Presumed diagnoses were deter- mined after correlating the radiologi- cal diagnosis with blood, CSF and non-neurological manifestations. The 2 patients with cerebral infarctions had meningitis. In 1 patient TB meningitis (TBM) was confirmed by culture positivity, and in the other cryptococcosis was identi- fied on India ink stains and confirmed by fungal cultures and latex antigen tests. Of the 17 patients with radiologi- cal diagnosis of presumed TB, 14 patients had active PTB (sputum AFB-positive), 1 patient had histolog- ically confirmed TB adenitis, 1 patient had PCP and 1 patient had no non- neurological disease. The CSF and blood studies in these patients showed no specific diagnostic abnormalities other than neurosyphilis in 1 patient. TB cultures were negative in all 17 patients. One patient had a positive TB PCR in the CSF. This patient had PTB and a solitary enhancing FBL. As indicated above, all 17 patients had increased basal meningeal enhance- ment. In this group correlating the radiological diagnosis with blood, CSF and non-neurological manifesta- tions did not alter the diagnosis. These patients were therefore treated with the 4-drug anti-TB regimen as first- line treatment. Five patients had radiological diagnoses of TB/toxoplasmosis (2 of whom had associated neurocysticer- cosis). In these 5 patients, 1 had posi- tive toxoplasma serology in the blood and CSF. This patient had multiple enhancing FBL (including lesions suggestive of neurocysticercosis) on CT scan with basal meningeal enhancement. The patient also had PTB. One patient had a positive blood toxoplasma serology, no PTB, and multiple enhancing FBL with no basal meningeal enhancement on CT scan. The 3 other patients had PTB, nega- tive toxoplasma serology, multiple enhancing FBL with no basal meningeal enhancement (including lesions suggestive of neurocysticerco- sis in 1 patient) on CT scan. In this group correlating the radiological diagnosis with blood, CSF and non- neurological manifestations influ- enced the presumed diagnosis. One of these 5 patients was therefore diag- nosed with isolated toxoplasmosis and treated with anti-toxoplasmosis treatment as first-line therapy. Of the remaining patients in this group, 2 were classed as TB, 1 as TB and neu- rocysticercosis, and 1 as TB, toxoplas- mosis and neurocysticercosis, and treated accordingly. The following presumed diag- noses were therefore made: 17 patients (53%) TB, 6 patients (19%) neurocysticercosis, 2 patients (6%) TB and neurocysticercosis, 2 patients (6%) multiple infarcts (1 TBM, 1 cryptococcal meningitis), 1 patient TB and neurosyphilis, 1 patient toxoplasmosis, 1 patient toxo- plasmosis, TB and neurocysticercosis, 1 patient PML, and 1 patient PCNSL. In the patients with multiple infec- tions as diagnoses, these were based on positive serology in the case of neurosyphilis, positive serology and CT scan features in the case of neuro- cysticercosis, basal enhancement on CT scan with non-neurological TB in the case of TB, and positive serology in the case of toxoplasmosis. ORIGINAL ARTICLE 8 SA JOURNAL OF RADIOLOGY • July 2005 Fig.1. CT scan showing basal meningeal enhancement in a patient with presumed TB. 5-13pgs 9/28/05 9:20 AM Page 8 Treatment was instituted for each infection at the outset. Response to empirical treat- ment initiated The following overall response to treatment was obtained (Table III): (i) good response 22 patients (69%); (ii) poor response 3 patients (9%); and (iii) demised 7 patients (22%). In the patients with a presumed diagnosis of TB and treated only with anti-TB treatment (17 patients): 12 patients had a good response (Fig. 2) (CD4+ counts 21 to 610 cells/mm3), 3 patients had a poor response (CD4+ counts 1, 55, and 180 cells/mm3), and 2 patients demised (CD4+ counts 6, 14 cells/mm3). The 6 patients with neurocysticer- cosis (CD4+ counts 106 - 768 cells/mm3) responded well to treat- ment. The 2 patients with infarctions as FBL (CD4+ counts of 96 and 240 cells/mm3) demised. The patient treated as isolated tox- oplasmosis (CD4+ count 8 cells/mm3) demised. In the mixed infection group (2 TB and neurocysticercosis, 1 TB and neurosyphilis, 1 TB, toxoplasmosis and neurocysticercosis) response to treatment was good in all 4 cases (CD4+ counts of 30 - 104 cells/mm3). The patients diagnosed with PCNSL (CD4+ count 22 cells/mm3) and PML (CD4+ count 7 cells/mm3) demised. The patient with presumed PML had autopsy confirmation of this diagnosis. The 3 poor responders to first-line TB treatment (all presumed TB) received additional anti-toxoplasmo- sis treatment and broad-spectrum antibiotics. There was no clinical or radiological change in spite of these added treatments during the follow- up period. Discussion The FBLs seen in our patients were determined to be almost exclusively infectious in nature. Of the 32 patients only 1 patient was presumed to have a non-infectious aetiology (PCNSL). Four patients (12%) were diagnosed with dual/multiple infections. Dual/multiple CNS infections are not common in HIV but have been described.13 We determined the principal pre- sumed cause of FBL in our patients to be TB. The key features differentiating suspected TB from other infectious aetiologies were: concurrent presence of PTB or other non-neurological TB, and basal meningeal enhancement. The FBLs in the patients with pre- sumed TB were predominantly corti- cal, with thick irregular walls, hypo- dense centres, and rim enhancement. Despite detailed analysis, of all the parameters used the only possible dif- ferentiating radiological feature was the presence of basal meningeal enhancement in TB (Table II). Of the 22 patients presumed to have TB (alone or in combination with other infections), basal meningeal enhance- ment was present in 16 patients (72%). The patients with presumed toxoplasmosis and neurocysticercosis had no basal meningeal enhance- ment. Patients with multiple infec- tions who had basal meningeal enhancement all had TB as one of the presumed infections. The only patient with basal meningeal enhancement whose presumed diagnosis was not TB, had cryptococcal meningitis with cortical infarctions as the FBL. TB cul- tures were not helpful in the acute management of our patients. Only 1 of our patients had a positive CSF TB culture. This was the patient with TBM and cortical infarctions as the FBL. CSF TB PCR and ELISA were also of no value in our study, being positive in only 1 patient respectively. In terms of non-neurological TB, this was diagnosed in 21 (66%) of our patients, of whom 19 had active PTB. TB is endemic in our population.8,11 In the Gauteng province of South Africa in 1999, the incidence of TB was 270/100 000 population. In 2001 the incidence was 315/100 000. PTB accounts for 75% and extra-pul- monary TB accounts for 25% of cases (Gauteng Department of Health — ORIGINAL ARTICLE 9 SA JOURNAL OF RADIOLOGY • July 2005 Fig. 2a and b. CT scans of a patient with pre- sumed TB at presentation (top) and on anti- TB treatment after 6 weeks (bottom). 5-13pgs 9/28/05 9:20 AM Page 9 unpublished data obtained from Dr Riana Louw, with permission). Sixty- three per cent of our patients had PTB. It is therefore not surprising that our presumed diagnoses showed a bias towards TB. With regard to toxoplasmosis we found that only 2 of our patients had positive serology (1 in serum, 1 in CSF and serum). Prevalence figures for toxoplasmosis in our region and South Africa as a whole have been published. In Gauteng the prevalence amongst blacks is 29%,14 and in KwaZulu-Natal the prevalence is 46% in black pregnant women, the highest in the country.15 It may be, as is wide- ly recognised, that serological tests are falsely negative in patients with advanced immunosuppression. However, active non-neurological toxoplasmosis has not been described in our population. We were unable to find data for toxoplasma retinitis, tox- oplasma septicaemia, and toxoplasma pneumonitis in our adult population (with or without HIV). None of our patients had evidence of toxoplasmo- sis outside the nervous system. The KwaZulu-Natal study found toxoplas- mosis as the main cause of HIV asso- ciated FBL.3 In another South African series of 38 HIV-infected patients with meningitis, toxoplasma IgG was positive in 14 patients (37%).16 Two of these 14 patients also had FBL, both of which were TB (one confirmed histo- logically) and responded to TB treat- ment.16 The only other study from this country on HIV-associated FBL con- curred with our findings that TB is the commonest cause of HIV-associated ORIGINAL ARTICLE 10 SA JOURNAL OF RADIOLOGY • July 2005 Table I. Focal brain lesions (FBLs) – clinical and laboratory findings with presumed cause/s Neurological Non-neurological CD4+ count Patient data examination illnesses CSF results (cells/mm3) Presumed cause N = 32 Focal signs 66% PTB 60% Elevated protein 59% 72%: < 200 53% TB M/F ratio 1.5: 1 Normal 28% PTB+TB breast Pleocytosis 9% 13%: 200 - 500 19% NCC (1 patient) Mean age 33.5 yrs Dementia 19% TB lymphadenitis TB PCR/Elisa + 6% 15%: > 500 3% Toxoplasmosis (1 patient) Encephalopathy 6% PCP (1patient) Cryptococcal Ag + 3% 3% PML Syphilis (11 patients) Cysticercosis Elisa + 21% 3% PCNSL Syphilis + 3% 6% Infarcts Toxoplasma + 3% 13% Mixed Normal 22% CSF = cerebrospinal fluid, TB = tuberculosis, PTB = pulmonary tuberculosis, PCP = Pneumocystis carinii pneumonia, NCC = neurocysticercosis, PCNSL = primary central nervous system lymphoma, PML = progressive multifocal leucoencephalopathy, PCR = polymerase chain reaction, ELISA = enzyme linked immunosorbent assay. Table II. Imaging characteristics of focal brain lesions (FBLs) FBL enhancement Basal meningeal N Type of lesions Site of lesions Wall definition Density relative to brain characteristics enhancement 90 Solitary 38% Cortical 79% Regular Hyperdense 6% Ring 63% Present 53% Multiple 62% Subcortical 11% Thick 0% Isodense 9% Nodular 6% Absent 47% Basal ganglia 7% Thin 100% Hypodense 85% Lobular 3% BS/cerebellar 3% Irregular Gyriform 5% Thick 61% Mixed 21% Thin 39% Nil 2% N = number of FBLs. 5-13pgs 9/28/05 9:20 AM Page 10 FBL.4 Further studies are required to clarify this issue. In our patients the presence of positive toxoplasma serol- ogy was used as an indication for tox- oplasmosis treatment. Neurocysticercosis, alone or in combination with other infections occurred in 9 patients (28%). This may be a chance association of endemic infections.17 In a case report of 4 patients from Zimbabwe, the occurrence of neurocysticercosis in HIV-positive patients was likewise suggested to be due to chance associa- tion in endemic regions or due to the effect of HIV on the host immune response to cysticercosis.17 Non-spe- cific host factors (innate resistance) together with acquired immunity are known to have an effect on the out- come of the primary infection in experimental models of cysticercosis.17 Host immunity is also an important factor in limiting the occurrence of cysticercosis in humans.17 The immunodeficient state accompanying HIV infection might therefore increase the frequency and severity of neurocysticercosis.17 In our series the mean CD4+ count in patients with neurocysticercosis alone was 509 cells/mm3. These patients were there- fore not immunocompromised. This implies a chance association. Further studies in respect of this association will be needed. In contradistinction to the histopathologically based American guidelines,1 the initial treatment in the majority of our patients, where diag- nosis (presumed) was based on collat- ing clinical information (as described above), was directed against TB. Only 1 patient had initial toxoplasmosis treatment alone. Patients with multi- ple infections were treated for these different infections. Sixty-nine per cent of patients improved clinically and radiologically. These patients had a mean CD4+ count of 188 cells/mm3. Fifty-five per cent of these good responders were diagnosed with TB. Twenty-seven per cent of the good responders had neurocysticercosis. The remainder had dual/multiple infections associated with TB. There were 3 poor responders (9%). The mean CD4+ count in these patients was 80 cells/mm3. All 3 were diag- nosed with TB. Seven of the 32 patients demised. These patients had a mean CD4+ count of 28 cells/mm3. In these patients the diagnoses were TB, toxoplasmosis, PML, PCNSL, and cryptococcosis. The good responders had higher CD4+ counts than those who responded poorly or demised. There was no similar correlation with blood HIV viral loads. This may imply that our diagnoses were more accurate in patients with CD4+ counts in excess of 100 cells/mm3 but that they were less accurate when the CD4+ counts were less than 100 cells/mm3. The patients who demised or responded poorly did not improve clinically or radiologically once additional treat- ments were instituted as described above. The patients who demised deteriorated rapidly, with death occurring within a few days of presen- tation to hospital. This may therefore also imply that the poor responders ORIGINAL ARTICLE 11 SA JOURNAL OF RADIOLOGY • July 2005 Table III. Treatment outcomes of focal brain lesions (FBLs) CD4+ count CD4+ count HIV viral loads Patients (N (%)) Response (mean - cells/mm3) Presumed cause (range - cells/mm3) (range - copies/ml) 22 (69%) Good 188 TB 55% 21 - 610 76 000 - > 750 000 NCC 27% 107 - 768 10 751 - 305 123 TB + other 18% 63 - 238 288 281 - > 750 000 3 (9%) Poor 80 TB (100%) 1 - 180 647 696 - > 750 000 7 (22%) Demised 28 TB 44% 6 - 40 242 000 - > 750 000 Toxoplasmosis 14% 8 ND PCNSL 14% 22 > 750 000 PML 14% 7 > 750 000 Cryptococcosis 14% 96 228 668 TB = tuberculosis, NCC = neurocysticercosis, PCNSL = primary central nervous system lymphoma, PML = progressive multifocal leucoencephalopathy, ND = not done. 5-13pgs 9/28/05 9:20 AM Page 11 ORIGINAL ARTICLE 12 SA JOURNAL OF RADIOLOGY • July 2005 and those who demised did so from advanced immune deficiency rather than incorrect diagnosis and treat- ment of the FBL. A limitation of our study is the small number of patients and the lack of long-term follow-up. The patients who improved were discharged and continued treatment as outpatients. After a few consultations most were lost to further follow-up. Conclusions Our data demonstrate that apply- ing the ‘developed nations guidelines for HIV associated FBL’ to a develop- ing region could have a deleterious effect on outcome. This is largely because non-infectious aetiologies, namely PCNSL are uncommon and TE is not necessarily the most fre- quent infectious aetiology. We agree with the British proposal that brain biopsies are not necessary for treat- ment. The AAN guidelines are not possible to implement in developing regions, because of financial con- straints and the limited availability of neurosurgical and neuropathological resources and services. In developing countries HIV-posi- tive patients who present with FBL should be treated initially with med- ication specific to the infection that is endemic to that population. In areas where TB is endemic our results indi- cate that anti-TB treatment should be Fig.3. Algorithm for the initial assessment of HIV-associated focal brain lesions (FBLs) in developing regions.( TB =tuberculosis, NCC = neurocysticer- cosis, TOXO = toxoplasmosis, PCNSL = primary central nervous system lymphoma, PML = progressive multifocal leucoencephalopathy, + = positive, - = negative.) 5-13pgs 9/28/05 9:20 AM Page 12 the initial treatment of known or sus- pected HIV-associated FBL. Further studies of this nature from other developing regions are needed to vali- date our approach especially with respect to the role of the endemic infection, which is treatable and would influence outcome and sur- vival. Perhaps the most important con- tribution of our study is that we show that detailed laboratory investigations (such as HIV viral loads, viral cul- tures, TB cultures, TB ELISA, TB PCR, and even CD4 counts) are not necessary to arrive at a reasonable pre- sumed diagnosis of a FBL in HIV- positive patients in order to institute effective treatment. CT scans of the FBL, with the exception of neurocys- ticercosis and possibly PML, were found to be nonspecific. The most useful parameters that assisted us were non-neurological illness, toxoplasma serology, and meningeal enhance- ment on CT scans. Algorithm (Fig. 3) On the basis of our results we pro- pose the following approach to the presumed diagnosis and thereby ini- tial treatment of HIV-associated FBL in developing regions. The important parameters as discussed above are meningeal enhancement on CT scan, presence of non-neurological TB, and toxoplasma serology. The FBLs are first categorised into solid or cystic. The cystic FBLs are indicative of neurocysticercosis (except in rare cases, e.g. hydatid). The solid FBLs are subdivided into enhancing or non-enhancing lesions. Lack of enhancement suggests PML. Enhancing FBLs are divided into those with and those without increased basal meningeal enhance- ment. The presence of increased basal meningeal enhancement is suggestive of TB. FBLs without increased meningeal enhancement are separat- ed on the basis of accompanying non- neurological TB and toxoplasma serology (see algorithm). Initial treat- ment is instituted accordingly. Patients are monitored clinically. Good responders are maintained on this treatment. Poor responders are reassessed with CT scans. Presumed diagnosis and treatment is appropri- ately adjusted, e.g. addition of anti- toxoplasma treatment, antibacterial treatment (pyogenic abscess). A limitation of this approach is the requirement of a CT scan for evalua- tion. In developing regions where this is not available, it might be best advised to treat suspected FBL in HIV patients for TB (in endemic areas) and toxoplasmosis. References 1. Quality Standards Subcommittee of the American Academy of Neurology. Evaluation and management of intracranial mass lesions in AIDS. Neurol 1998; 50: 21 - 26. 2. Sadler M, Brink NS, Gazzard BG. Management of intracerebral lesions in patients with HIV: a retrospective study with discussion of diagnos- tic problems. QJM 1998; 91: 205 - 217. 3. Bhigjee AJ, Naidoo K, Patel VB, Govender D. Intracranial mass lesions in HIV-positive patients – The KwaZulu/Natal experience. S Afr Med J 1999; 89: 1284-1288. 4. Smego RA, Orlovic D, Wadula J, Modi G. A diagnostic and therapeutic algorithm for CNS mass lesions in HIV/AIDS. S Afr J Epidemiol Infect 2000; 15: 7-13. 5. Trujillo JR, Garcia-Ramos G, Novak IS, Rivera VM, Huerta E, Essex M. Neurologic manifesta- tions of AIDS: a comparative study of two pop- ulations from Mexico and the United States. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 8: 23-29. 6. Lucas SB, Hounnou A, Peacock C, et al. The mortality and pathology of HIV in a west African city. AIDS 1993; 7: 1569-1579. 7. Modi M, Modi G. New onset seizures in HIV infected patients – a review and guide to man- agement. S Afr Med J 2001; 91: 1025-1026. 8. Karstaedt AS, Valtchanova S, Barriere R, Crewe- Brown HH. Tuberculous meningitis in South African urban adults. QJM 1998; 91: 743-747. 9. Santosh V, Shankar SK, Das S, et al. Pathological lesions in HIV positive patients. Indian J Med Res 1995; 101: 134-141. 10. Mathew MJ, Chandy MJ. Central nervous sys- tem toxoplasmosis in acquired immunodefi- ciency syndrome: An emerging disease in India. Neurol India 1999; 47: 182-187. 11. Raviglione MC. The TB epidemic from 1992 to 2002. Tuberculosis 2003; 83: 4-14. 12. Centers for Disease Control. 1993 revised classi- fications system for HIV infection and expand- ed surveillance case definition for AIDS among adolescents and adults. MMWR CDC Surveill Summ. 1993: 1-19. 13. Silber E, Sonnenberg P, Koornhof HJ, Morris L, Saffer D. Dual infective pathology in patients with cryptococcal meningitis. Neurology 1998; 51: 1213-1215. 14. Mason PR, Jacobs MR, Fripp PJ. Serological survey of toxoplasmosis in the Transvaal. S Afr Med J 1974; 48: 1707-1709. 15. Schneider E, Schutte CHJ, BommerW. The prevalence of Toxoplasma gondii infection in women of different ethnic groups in Natal, South Africa. S Afr J Epidemiol Inf 1992; 7: 41- 45. 16. Sonnenberg P, Silber E, Jentsch U. Toxoplasmosis and HIV infection in Southern Africa. S Afr J Epidemiol Inf 1998; 13: 104-106. 17. Thornton CA, Houston S, Latif AS. Neurocysticercosis and human immunodefi- ciency virus infection – A possible association. Arch Neurol 1992; 49: 963-965. ORIGINAL ARTICLE 13 SA JOURNAL OF RADIOLOGY • July 2005 Reprinted from QJM 2004; 97: 413-421, with permission. 5-13pgs 9/28/05 9:20 AM Page 13